1
0
mirror of https://github.com/samba-team/samba.git synced 2025-02-03 13:47:25 +03:00

16924 Commits

Author SHA1 Message Date
James Peach
5712644fcc r15928: When we search for clock_gettime, we might find it in librt, but librt
might pull in libpthread. This is quite bad, firstly because it can
cause oplock signals on Linux to go wonky, and secondly because merely
linking with pthreads can cause performance degradations due to implicit
locking requirements.

The solution is to only search for clock_gettime if --with-profiling-data
was specified. If we do end up searching for it, then we test whether
linking with librt pulled in libpthread, and we only allow the definition
for clock_gettime to succeed if libpthread was NOT linked in.

Problem reported by Thomas Bork and diagnosed by Volker Lendecke.
2007-10-10 11:17:13 -05:00
Volker Lendecke
53bd3fa30a r15924: Ok, the Solaris fix broke AIX. Next try :-)
Volker
2007-10-10 11:17:13 -05:00
Volker Lendecke
c9542c64ce r15918: Attempt to bring Solaris into make test. The Solaris shell does not like the
isolated !.

Metze, can you check this?

Thanks,

Volker
2007-10-10 11:17:13 -05:00
Volker Lendecke
17f701ab4f r15917: Activate RPC-AUTHCONTEXT in "make test" 2007-10-10 11:17:13 -05:00
Volker Lendecke
c24bfdce62 r15911: Make us survive rpc-authcontext committed next 2007-10-10 11:17:13 -05:00
Volker Lendecke
09f3c7a86f r15910: vfs_full_audit does not need current_user 2007-10-10 11:17:13 -05:00
Volker Lendecke
4dd8694a25 r15909: Implement recycle:subdir_mode 2007-10-10 11:17:12 -05:00
Gerald Carter
0a81af4fef r15906: smbpasswd help text for -W option (patch from Aruna Prabakar <aruna.prabakar@hp.com> 2007-10-10 11:17:12 -05:00
Volker Lendecke
b1244e7906 r15904: This does two things:
Fix more potential segfaults when something on our way to a DC connection
fails.

We can not continue if dcip_to_name() fails. With

192.168.234.100 nt4pdc
192.168.234.100 windows#1c
192.168.234.100 windows#1b

in the lmhosts file when nt4pdc is rebooted, we do find the DC's IP address,
we can connect to TCP 139 while it is booting but anything else fails. So we
fall back to put the IP address into domain->dcname. When the DC is fully up
later on we try to do the auth2 against \\192.168.234.100 which gives
INVALID_COMPUTER_NAME. And we never get out of this loop again.

Fix this.

Jerry, maybe you can take a look.

Thanks,

Volker
2007-10-10 11:17:12 -05:00
Jeremy Allison
4962548dfe r15895: Ensure all new rid allocation goes through
the same function (deals with races).
Jeremy.
2007-10-10 11:17:12 -05:00
Jeremy Allison
a8eb1186a1 r15890: Use correct enum type (bug #3722) from Jason Mader <jason@ncac.gwu.edu>.
Jeremy.
2007-10-10 11:17:12 -05:00
Jeremy Allison
9890a31c5f r15888: Fix bug #3804 from jason@ncac.gwu.edu
Invalid comparisons.
Jeremy.
2007-10-10 11:17:12 -05:00
Jeremy Allison
c66620770d r15887: Ensure we use sys_write so we're not interrupted.
Jeremy.
2007-10-10 11:17:12 -05:00
Jeremy Allison
1d9dbe3b66 r15884: Fix bug #3803 from jason@ncac.gwu.edu - bad comparison.
Jeremy.
2007-10-10 11:17:11 -05:00
Jim McDonough
de836d2077 r15873: Use short signal names to placate ksh trap. bash seems to accept either
and even uses them on the manpage.  this should now enable make test on
AIX.
2007-10-10 11:17:11 -05:00
Jim McDonough
0ee78d96d9 r15872: Take one step toward getting this working on AIX. the [-z ] stuff
doesn't work there.
2007-10-10 11:17:11 -05:00
Gerald Carter
ece5fe3e78 r15864: build fixes for IRIX 6.4 in the build farm; only enable the C++ reserved word check when we selecte --enable-developer 2007-10-10 11:17:11 -05:00
Gerald Carter
094443ed6c r15863: trying to fix a compile issue on HP-UX 11.x caused by conflicts of int16 and int32 definitions in internal and system headers 2007-10-10 11:17:11 -05:00
James Peach
bf3b71c845 r15848: Introduce commandline options to set the remainder of the parameters in
dynconfig.c. This is mainly useful for test harness scripts, hence the
lack of short options.
2007-10-10 11:17:11 -05:00
Volker Lendecke
c8a8204c74 r15845: Ok. This was a tough one. If for some reason the tconX fails towards a domain
controller the next time we connect this child ran into a segfault because it
tried to reference a half-baked connection.

Volker
2007-10-10 11:17:11 -05:00
Gerald Carter
39f7ff75a7 r15842: patch from volker to instruct winbindd to find a trusted DC on its own when runing on a Samba DC (since we don't implement the getdcname() call that well 2007-10-10 11:17:10 -05:00
Jeremy Allison
f6c110ddb8 r15838: Back-port tridge's talloc fixes (r15824, r15828) from Samba4.
Jeremy.
2007-10-10 11:17:10 -05:00
Lars Müller
7c375fd540 r15822: Add suggestion made by Ralf Haferkamp. 2007-10-10 11:17:10 -05:00
Volker Lendecke
72f103708d r15817: Remove some unused code 2007-10-10 11:17:10 -05:00
Volker Lendecke
b013b6908d r15755: Fix Coverity bug # 294. Apparently password can be NULL, but cli_session_setup
derefences it.

Volker
2007-10-10 11:17:10 -05:00
Stefan Metzmacher
8709bba24a r15735: also export PREFIX as testit from test_functions.sh needs it
metze
2007-10-10 11:17:10 -05:00
Stefan Metzmacher
e5a8553a57 r15732: samba3 now passes BASE-DENY1
metze
2007-10-10 11:17:09 -05:00
Jeremy Allison
5137006233 r15729: Second part of Aleksey Fedoseev <fedoseev@ru.ibm.com> patch.
Jeremy.
2007-10-10 11:17:09 -05:00
Jeremy Allison
9f5c95c418 r15727: Fix msgtest - noticed by Aleksey Fedoseev <fedoseev@ru.ibm.com>.
Jeremy.
2007-10-10 11:17:09 -05:00
Gerald Carter
94e0bc5013 r15717: BUG 3711: patch from Leonid Kabanov <lkabanov@mail.ru> for some solaris sheel scripting bugs with 'make test' 2007-10-10 11:17:09 -05:00
Gerald Carter
466478f07e r15712: BUG 3435: patch from volker to fix 'msdfs root = yes' in [homes] 2007-10-10 11:17:09 -05:00
Volker Lendecke
e4a2cb4b91 r15705: Fix bug number 3788. Thanks to Jeff Wright.
Volker
2007-10-10 11:17:09 -05:00
Günther Deschner
6cfc65ea20 r15704: Prefer LDAP error codes in ads_search_retry_sid().
Guenther
2007-10-10 11:17:08 -05:00
Günther Deschner
741602e03a r15703: Fix d_printf call.
Guenther
2007-10-10 11:17:08 -05:00
Gerald Carter
ad1f947625 r15701: change 'net ads leave' to disable the machine account in the domain (since removal implies greater permissions that Windows clients require) 2007-10-10 11:17:08 -05:00
Jeremy Allison
e5787cf75b r15700: Make nmbd udp sockets non-blocking to prevent problem
with select returning true but no data being available.
Fix for bug #3779.
Jeremy.
2007-10-10 11:17:08 -05:00
Günther Deschner
7d766b5505 r15698: An attempt to make the winbind lookup_usergroups() call in security=ads
more scalable:

The most efficient way is to use the "tokenGroups" attribute which gives
the nested group membership. As this attribute can not always be
retrieved when binding with the machine account (the only garanteed way
to get the tokenGroups I could find is when the machine account is a
member of the "Pre Win2k Access" builtin group).

Our current fallback when "tokenGroups" failed is looking for all groups
where the userdn was in the "member" attribute. This behaves not very
well in very large AD domains.

The patch first tries the "memberOf" attribute on the user's dn in that
case and directly retrieves the group's sids by using the LDAP Extended
DN control from the user's object.

The way to pass down the control to the ldap search call is rather
painfull and probably will be rearranged later on.

Successfully tested on win2k sp0, win2k sp4, wink3 sp1 and win2k3 r2.

Guenther
2007-10-10 11:17:08 -05:00
Günther Deschner
52423e01dc r15697: I take no comments as no objections :)
Expand the "winbind nss info" to also take "rfc2307" to support the
plain posix attributes LDAP schema from win2k3-r2.

This work is based on patches from Howard Wilkinson and Bob Gautier
(and closes bug #3345).

Guenther
2007-10-10 11:17:08 -05:00
Günther Deschner
ec26c355b3 r15696: Free LDAP search result.
Guenther
2007-10-10 11:17:07 -05:00
Stefan Metzmacher
b79ceece95 r15691: - samba3 also pass RAW-CHKPATH, RAW-SEEK, RAW-SFILEINFO-BUG
- samba3 pass the limited version of RAW-LOCK
  (the lock cancel and error checking tests are skipped)

metze
2007-10-10 11:17:07 -05:00
Stefan Metzmacher
fc343051af r15686: - start testing the RAW-* tests in make test
- RAW-READ now passes samba3
- skip the rest for now

metze
2007-10-10 11:17:07 -05:00
Stefan Metzmacher
cd2fa1da6c r15682: fix option passing
metze
2007-10-10 11:17:07 -05:00
Gerald Carter
d77768cb23 r15681: fix segv in 'kinit && net ads join' 2007-10-10 11:17:07 -05:00
Gerald Carter
ae2bf464c4 r15680: use the user creds when calling net_set_machine_spn() rather than the machine creds (just like WinXP) 2007-10-10 11:17:07 -05:00
Gerald Carter
126ab8ccec r15679: fix some popt warnings 2007-10-10 11:17:06 -05:00
Jeremy Allison
ff3fe39b83 r15676: Fix meaningless debug statement from uninitialized variable.
Spotted by "John E. Malmberg" <wb8tyw@qsl.net>.
Jeremy.
2007-10-10 11:17:06 -05:00
Jeremy Allison
256ae3a16b r15675: Man pages say never look at the fd_set after a select
if it returned -1 (treat as undefined). Ensure we obey
this.
Jeremy.
2007-10-10 11:17:06 -05:00
Jeremy Allison
24aaa48677 r15672: Fix for bug #3783. winbindd_cm.c calls open_any_socket_out()
to make connections to ports 445 or 139 on the DC it's trying
to contact. It calls sys_select() on the non-blocking sockets,
not sys_select_intr(). This is a mistake (I believe) as it allows
a signal to early terminate the connection attempts - whereas
sys_select_intr() will ignore signals until we get back to
the main processing loop where they'll be handled correctly.
This change means winbindd_cm will not early terminate if it
gets a message whilst trying to connect to DC's.
Gunther, Volker and Jerry please review (but I think this
is correct).
Jeremy.
2007-10-10 11:17:06 -05:00
Jeremy Allison
31d16c434e r15670: Fix valgrind-spotted issue in BASE-DELETE test.
We were forgetting to increment after copying
the primary group gid.
Jeremy
2007-10-10 11:17:06 -05:00
Jeremy Allison
fd6fa1d4ea r15668: DOS or FCB opens share one share mode entry from different
fsp pointers. Ensure we cope with this to pass Samba4
DENY tests (we used to pass these, there must have been
a regression with newer code). We now pass them.
Jeremy
2007-10-10 11:17:06 -05:00